Researchers in the Netherlands are working on a prototype autonomous litter picking robot that uses AI to recognise and clean up waste.
Students at University of Technology Delft, studying Minor Robotics, assisted in the development of BeachBot. The litter-collecting rover picks up small items such as cigarette butts, single-use cutlery, or plastic bottle caps from beaches.
With the help of a game, the self-learning robot is trained using an image recognition algorithm. In the case of unknown waste, the robot places a photo with a GPS tag in an app. Users of the app can indicate what kind of waste it concerns, so that the robot can identify this in the future. In this way the robot learns to recognise more and more waste.
The team are working towards a proof-of-principle, but explained that more needs to be done at a policy level, and the public must also get involved.
Creator Edwin Bos, said: “It’s nice to develop a robot solution, but that’s not the solution to the wider problem.
“Behaviour needs to change, and our goal is to make people interact and engage with the robot to make it smarter, but also to learn about the impact of litter themselves.”
A BeachBot prototype has been deployed in several locations in the Netherlands and the entrepreneurs said they’re ready to move toward launching the product.
According to the project, the next challenge is to find the right business model to ensure BeachBot doesn’t just clean, but also educates the public and changes behaviours.
